Professional Documents
Culture Documents
IP Telephony from A to Z
Chapter 1
Chapters One to Eight E
ebook
Chapter 2
The Decision: Vendor Evaluation and Selection
Provides you with resources to help you evaluate and select IP telephony vendors.
Chapter 3
Planning: The Implementation Calendar
Provides you with a high-level timeline for the implementation, from research to actual deployment.
Chapter 4
Ensuring Reliability in IP Telephony
Covers varying IP telephony solution architectures, mean time between failure, mean time to repair,
network reliability, and application reliability.
Chapter 5
Handsets and Interfaces
Outlines the many benefits of todays’ well-designed and highly functional telephones.
Chapter 6
Security
Highlights the steps one should take to ensure IP telephony traffic is secure against outsiders and
unauthorized individuals.
Chapter 7
Mobility and Wireless
Highlights the requirements and methods for going wireless.
Chapter 8
Quality of Service
Covers Quality of Service (QoS) in detail, as well as your options in terms of circuit transports.
Then, it delves into the internal infrastructure and the entire process of applying QoS.
The Benefits of IP Telephony
1
Chapter 1
TABLE OF CONTENTS
This chapter highlights the benefits of IP telephony and discusses the costs
The Savings 1
IP Telephony Savings 2
The Costs 4
The Savings
When you consider what most businesses pay for long-distance, you wouldn’t see a huge need to move
to IP telephony, necessarily. Large corporations can be paying pennies per minute for long-distance
within the U.S. So while companies beyond North America may realize significant savings on toll
charges, these savings are not usually enough to convince a North American company to switch to IP
telephony.
Savings for most enterprise networks come from consolidating the voice and data network and using
fewer circuits from the public switched telephone network (PSTN). In addition to circuit cost savings,
as mentioned earlier, an IP infrastructure requires less time for moves, adds and changes (MACs) and
often eliminates the need to hire an outside vendor or service provider to handle them. Moving an IP
telephone station temporarily or permanently or adding a new user usually simply entails carrying out a
quick and simple GUI-based command. With traditional PBX systems, moving an employee can cost
hundreds of dollars in labor. In other words, with IP telephony, each user has their own IP phone profile
and the network doesn’t care where anybody is located at any particular time, so MACs are simply a
matter of conducting a few commands and can often be easily handled by the user.
With IP telephony, management savings are usually immediate since the information technology team
can support the voice network as well as the data network because they’re now one in the same. There
is no longer a need to have two teams of technical professionals to handle each entity, which adds up to
tremendous savings. Further savings are seen right away when an enterprise needs to make a change,
such as re-locating an office temporarily in the case of construction. The IT staff simply makes the
changes from anywhere on the network (or remotely if need be) and a new temporary office is up and
running without outside callers ever being the wiser.
Finally, infrastructure tools like physical ports are no longer needed for IP telephony because physical
circuit-switched ports aren’t necessary. An IP connected voice mail server is all that’s needed.
All of these cost savings are tremendously appealing characteristics of IP telephony. When you add to
them the features that are available for employees, call centers and receptionists, it quickly becomes
obvious that IP telephony is going to continue winning converts.
Another customer service feature available in IP telephony solutions is the hunt group. This feature makes
certain that all calls are answered by a live person rather than voice mail, which can be frustrating for
callers. With various hunt groups enabled, a call into an organization rings extensions in a specified se-
quence or rings multiple extensions at once (depending on the company’s preference), ensuring callers
reach the person they need without navigating through menus or being forced to wait in a queue.
Remote sites are also easy to bring online. With traditional PBX systems, adding a remote site often re-
quires adding a PBX extender, which can cost almost $1,000 per user for the equipment alone. With IP
telephony, again, a user can log in from anywhere and have all the same capabilities as if they were working
at headquarters or within the call center building. With IP telephony, to the outside world, it can seem as
though you have call center locations scattered around the globe to be available 24/7, when really you are
simply utilizing IP telephony features such as time-of-day routing and call forwarding to make sure calls are
answered quickly by a live human being; these people can be working out of geographically-dispersed
branch offices, at remote locations, or even at home. Callers always reach a qualified customer service
representative, regardless of what time it is. You are also able to manage peak calling times by having the
ability to add other employees, regardless of their location, to the call center to help meet the overflow
demand.
With IP telephony, users can also easily re-route their calls so that they are reached wherever they will be
working—they can make these changes themselves, without asking for IT assistance. This “find me” feature
also enhances customer service as well as productivity by ensuring a caller reaches the right person,
regardless of where he or she might be working. An employee can even program his or her extension to
ring based on status—ring through when he or she is in the office, forward to a cell phone when there is no
answer, or forward to a colleague when the line is busy.
IP telephony systems also allow organizations to implement skills-based routing, whereby calls are
routed via an automatic attendant (attendant prompts the caller to choose from a selection) to the
most appropriate agent based on criteria like language, experience, technical expertise, and other
details. Advanced features that most service providers charge for are also available “free” with IP
telephony, including three-way calling and a built-in conference call bridge. This can further aid in
customer service when more resources are required to fulfill a customer request or inquiry, and it also
allows conference call access by international parties, a feature most expensive conference call services
do not provide.
Finally, IP telephony enables self-service options. For instance, when a caller simply wants to find out
information about their own account, interactive voice response (IVR) within IP telephony systems
enable callers to securely access that information by providing specific information. This eliminates the
need for a call center agent to take time to answer a call, and it also eliminates the frustration that can
occur if a caller is put in queue on hold for the next available agent to find out information that is
readily available.
Soft phones further free people from their desks, delivering telephony capabilities to any PC. With calls
directed to a laptop and a headset plugged into the USB port, employees can work from anywhere using
their computer and its built-in microphone. Employees who travel a lot appreciate the power and
simplicity of a soft phone and customers appreciate not having to dial different numbers to reach
someone who is traveling.
Nemertes Research, which is one of the few research firms that focuses specifically on IP telephony,
suggests that you start the process by carefully assessing the size of your rollout. This consideration is not
dependent on company revenue but how many stations you need the solution to support. You will analyze
solutions for the time it takes to install these stations, and estimate your growth and how your particular
solution’s scalability will affect the deployment.
The Costs
Nemertes Research interviewed IT professionals from a wide variety of companies and analyzed four
leading vendors in specific areas, including total hardware costs, network upgrades, IP handsets, manage-
ment tools, and conferencing/collaborative applications. From these in-depth interview came a comprehen-
sive report entitled, “Convergence & Next-Generation WAN Technologies” (February 2006). This section
will look at some of the costs involved in an IP telephony solution deployment, as well as provide high-level
results of the interviews conducted.
$1,250
Mean Capital Cost Per User
$1,000
$750
$1,272
$1,220
$1,094
$500
$817
$729
$578
$434
$250
$308
0.00
Cisco Avaya Nortel ShoreTel
Vendor for Cost Analysis
The planning and design phase of any rollout is one of the most important. Consider your team and
think of how you will divide up responsibilities. Also, consider whether you will need to add to or
reduce your team size. For the implementation, decide on a few team leaders who will commit to
making themselves available in the off-hours when necessary until the deployment is complete. The
best solutions are easy to implement rather quickly and seamlessly, but you will still want some key
people available throughout the deployment.
Installation is the time it takes to physically deploy and configure the solution—it does not include
training. Again, consider carefully who is available to help with the installation, taking all things into
consideration such as work schedule flexibility, knowledge and expertise, and the ability to work under
pressure. Consider your business and determine the best time to deploy the solution and when it will be
easiest to switch over to the IP telephony solution.
Next up is troubleshooting—the time it takes to make changes immediately after the deployment until
it works properly. Who is going to be available throughout the deployment right up until the minute you
determine that everything is working perfectly? Consider the first few days and how you’ll staff the help
desk around the clock with people who are substantially knowledgeable about the infrastructure, the
configuration, and the features of the handsets.
Management is the next cost consideration. What are your staff members doing each day to support the
solution? Can things be handled in-house, without wasting time and money on an outside vendor or service
provider to handle personnel MACs? According to Nemertes Research, MACs become very easy with IP
telephony: Research participants estimate the time involved for an IP MAC at a mere 10 minutes or less,
compared to the 30 to 90 minutes required for a TDM MAC. This means that total cost savings, depending
on the average number of MACs at a given organization, can be significant.
Nemertes Research ultimately calculated the total cost of ownership (TCO) for IP telephony solutions from
leading vendors (see Figure 1.2 below). These numbers were calculated considering all of the costs listed
above. This gives you an overview of costs for each vendor’s solution based on the implementation size.
$3,500
$3,000
$2,500
$2,000
$3,384
$3,365
$1,500
$2,537
$1,970
$1,000
$1,127
$1,384
$1,223
$865
$500
0.00
< 1,000 units/station > 1,000 units/station
A Network World special report suggests that organizations should consider transitioning to IP
telephony when:
• They are using IP Centrex lines that will support phone and Internet service on the same network.
Moving to IP telephony will immediately reduce costs because these lines are so expensive.
• The organization is moving to a new building. Since the wiring does not yet exist, it’s simple to
create a consolidated data and voice network.
• They are coming to the end of a PBX lease agreement or the current phone system is outdated,
obsolete or unsupported by a vendor or service provider.
• The company has offices in different area codes and employees dial a lot of long-distance numbers.
The reduction in toll charges will be immediate and significant.
You will also want to consider IP telephony for your organization if:
• Your locations shift in size often
• Locations are added regularly
• You have a relatively small technology staff
• You use a great deal of outsourced telephony services that are beginning to add up
• Many of your employees frequently work remotely
Once you’ve evaluated your organization carefully, analyzing the costs of your current telephony
solution along with your employee productivity and customer service needs, and decided that
indeed, IP telephony is the way to go, the next chapter will help you with the vendor evaluation and
selection process.
Expectations 1
Choices, Choices 2
Vendors Analyzed 2
Avaya 2
Cisco 2
Nortel 3
ShoreTel 3
Seeing is Believing 5
Decision Factors 7
Expectations
InfoTech reports that enterprise decision makers generally have three main areas of expectation that
help them choose the right vendor. These are areas you’ll want to consider as you embark upon the
vendor evaluation phase.
Choices, Choices
In their 2006 report, “Convergence & Next-Generation WAN Technologies,” Nemertes Research provides a
comprehensive and unbiased look at what organizations are doing specifically in terms of which vendors
they choose. It is an independent and impartial report that translates mountains of data into succinct
information organizations can use for convergence planning. The firm collected information by conducting
in-depth interviews with IT professionals from a wide variety of companies of various sizes spanning many
industries. While the industries varied greatly, all of the respondents had a similar interest in IP telephony
and were committed to making technology investments that enhance productivity and the bottom line and
prepare their organizations for the future. Nemertes Research presents an overview of how the respondents
have assessed IP telephony solutions and how they eventually selected a system vendor. Included in the
report are recommendations about which vendors to consider, including a complete IP telephony system
vendor analysis, how to thoroughly evaluate all of the solutions available, how to plan for convergence, and
how to actually conduct the rollout.
Vendors Analyzed
Organizations in the past have had few vendors to choose from. According to Nemertes Research, today
there are more than 25 vendors and carriers out there to meet IP telephony needs. The increase in compe-
tition means more innovation and better products from a wider selection of companies. Nemertes Research
analysts established that the most frequently evaluated IP telephony system vendors today are: Avaya,
Cisco, Nortel and ShoreTel. The following section will highlight each of those vendors, but keep in mind
that there are at least a dozen more to evaluate, depending on the size and particular needs of your
organization.
Avaya
Avaya offers IP telephony solutions with its IP Office and MultiVantage solutions, which include IP tele-
phones, as well as voice switches, media gateways, communication servers, wireless telephones, communi-
cation applications, and more. According to Nemertes Research, Avaya’s key strengths are its product
features, technology, and overall performance, while weaknesses, according to respondents, fall in the areas
of customer service, ease of use (installation and troubleshooting), management tools, and VAR expertise.
Cisco
Cisco is a recognized network infrastructure equipment leader and offers IP telephony solutions under its
Unified Communications family. Products include switches, telephones, communication applications, and
more. Nemertes Research notes that Cisco’s overall performance and technology areas have been rising
steadily, according to respondents, while product features have left much to be desired. However, many
networks are built on Cisco networking equipment and it would be hasty to overlook the company during
an IP telephony vendor review.
ShoreTel
ShoreTel offers end-to-end IP telephony solutions including its ShoreGear voice switches and
ShorePhone IP telephones, as well as communication applications, call center functionality, and more.
ShoreTel scored highest in all categories studied by Nemertes Research. Four specific areas in which
the company excels are value, technology, ease of installation and troubleshooting, and performance.
The company’s areas for improvement included management tools, solution experience and VAR
expertise.
1. Select a project leader. This person should be experienced in networking and IP telephony, if
possible, and should be able to answer basic technical questions related to the technology, if not the
specific vendor solutions.
2. Assess what you need from the IP telephony solution.
• Evaluate the current situation, including costs, etc.
• Identify key goals.
• Review most common product capabilities and decide on the importance of them.
• Determine if there will be training required.
• Estimate the cost of the project.
3. Record your requirements, goals, and recommendations in a tentative plan.
4. Present your plan to the appropriate organizational leaders (executive management, financial
department, etc.). Get their input before writing the proposal.
5. Write the proposal. A typical proposal contains:
6. Submit the proposal to the vendors you’ve selected in your long list. Your integration partner or
consultant, if you have one, can help you with this process, or simply e-mail or fax it to your vendor
list.
Step 1 In Theory
Ranking This section is where you will simply
0=unsatisfactory note whether the vendor offers specific
10=excellent capabilities
Figure 2.1 Sample weighted worksheet for vendor evaluation—not an exhaustive list.
Next, ask to see a demo and request a sample set-up to test the solution in your office so you can revise
the score based on actual experience. Once you have seen a demo or tested the solution, revise your
weighted worksheet to reflect your actual experience. See Figure 2.2 for the revised worksheet and
score.
Call forwarding No 0 10 0 0 10 0
Caller ID Yes 10 10 100 8 10 80
4 or 5-digit dialing across locations No 0 10 0 0 10 0
Figure 2.2 Revised sample weighted worksheet for vendor evaluation, with experiential
scores—not an exhaustive list.
Easy scalability
Make sure that the vendor you choose knows exactly how you will need to scale the system for your
specific needs. For instance, if your organization often grows and shrinks during different times of the
year or in some other cyclical manner, ask how new users would be added to support your growth
needs. Will new hardware need to be added and removed each time you grow and shrink? Or will the
system support your needs up to a certain point, regardless of how many times you change size?
Multi-vendor interoperability
Some vendors are known for requiring a full infrastructure overhaul to accommodate the new IP
telephony system. Be certain that you can use your existing network equipment with the new
solution, and make sure that when you add new gear, you can do so without needing to consider the
IP telephony system. IP telephony is only beneficial if it’s truly part of the network and it doesn’t bring
you new headaches or worries further down the line.
Ease of implementation/management/maintenance
IP telephony systems should make life easier for the IT team, not more difficult. Because the new
system works on the existing network, everything is managed similarly. If management of the
$3,500
$3,000
$2,500
$2,000
$3,384
$3,365
$1,500
$2,537
$1,970
$1,000
$1,127
$1,384
$1,223
$865
$500
0.00
< 1,000 units/station > 1,000 units/station
The next chapter will explore the IP telephony implementation from beginning to end, starting with research
and vendor evaluation and ending with the actual deployment, and will include a helpful timeline for you to
use.
Planning: The
Implementation Calendar 3
Chapter 3
The following resources can be helpful in your search for IP telephony information.
• CIO Magazine (www.cio.com)
• Network Computing (http://www.networkcomputing.com)
• Network World (www.networkworld.com)
• VoIP Magazine (www.voip-magazine.com)
• ComputerWorld (www.computerworld.com)
Trade magazines and their online counterparts do cover vendors, of course, but you can find unbiased
technology primers and overviews. It’s also helpful to read customer case studies about deployments to
learn about the experiences of those companies that have deployed IP telephony. Read case studies for
technology tips first, vendor specifics second.
You’ll read about each solution with your own organization in mind. Jot down questions as you click
through vendor web pages. You may get the answer to the question quickly, or it may remain on your list
until you eventually meet with the vendor. If your organization has many offices across the United States,
for instance, look at solution descriptions with scalability, flexibility, and ease of deployment mentioned
early. If your organization rarely changes in size and has a limited number of telephony requirements, look
for solutions that offer the basics at a very affordable price point.
Next, create a checklist or table with some common features. For instance, most IP telephony solutions
offer standard features like caller ID and three- or four-digit dialing. As you exhaust the common feature
list, start adding unique features that matter to your organization. Learn (or try to learn) what differenti-
ates each vendor you’re considering. If you save the differentiation for the vendor presentation, you likely
will get a skewed answer to the question, “What makes your solution different and superior?” This checklist
is just the beginning and you won’t do anything with it until the RFP phase.
Read articles about each vendor and mark items off your checklist as you determine what each offers. Start
with articles that the vendor links to (usually found under headings like “press coverage,” “news coverage,”
“case studies,” “success stories,” and “customer solutions” on the website). However, vendors obviously
will only highlight their true success stories. Use an Internet search engine to do a little sleuthing your-
self—you may find three or four stories about users’ unhappiness with a certain vendor. Dig for the dirt.
Use all of this information for your checklist and research notes.
1. Determine your business requirements. How will the system be used? How many calls per month
(or day) are made out of your office? Are those calls to customers or internal employees? How
many offices will you have on a system? Are there remote offices to consider?
2. Look at your LAN. What equipment are you using? Do you have an up-to-date network diagram?
Is the equipment current or outdated? Are you using Virtual LANs (VLANs) for security or perfor-
mance issues? VLANs improve voice quality by prioritizing voice traffic.
3. Assess your WAN. How much WAN bandwidth do you have between offices? How many home or
remote offices do you have and will you need dedicated circuits or will DSL suffice? Consider
whether managed IP services are a fit for your organization as an alternative to traditional dedicated
circuits.
Once you’ve collected all of the information and carefully evaluated your short list of vendors, think care-
fully about your organization’s priorities in general and start talking to customers. Be sure you get customer
references that have similar networks and similar business requirements to your own organization. Again,
ask to speak with recent customers: It’s easy to give you a list of happy customers. Ask for a list of the most
recent customers signed on—within the last three months, for instance—and call them about their experience.
In order to achieve toll-quality voice, you need to deploy IP telephony over a properly architected
network infrastructure - i.e., it has to provide sufficient throughput and meet latency, jitter and packet
loss requirements.
Throughput: How much bandwidth you need depends on the how many simultaneous calls your organiza-
tion has going on, the voice encoding scheme used in the IP handset or soft phone, and the signaling
overhead.
Latency and Jitter: Latency is the time it takes for a caller’s voice to be transported (packetized, sent
over the network, de-packetized, replayed) to the other individual. Distance and lower-speed circuits can
cause delay. Latency that’s too high interrupts the natural conversation flow (you may have spoken with
someone using VoIP - you think they have stopped talking but they haven’t-that’s latency). Latency cannot
exceed 100 milliseconds one way for toll-quality voice. Acceptable quality voice can go up to 150 millisec-
onds and participants can still carry on a decent conversation.
A thorough assessment uses active application traffic across the LAN and WAN in order to reveal what’s
going to happen when IP telephony is introduced into the mix. Test agents send a variety of network
traffic packets - using different application protocols, packet size, packet spacing and quality of service
(QoS) levels. The tests simulate the various types of IP telephony traffic that are likely to occur on a
live network. In addition to measuring peer-to-peer traffic, the agents can also generate real-time client
transactions against production servers, including communication with IP PBX servers. This compre-
hensive approach enables the test engineer to pinpoint the source of potential problems and make
recommendations for resolution, thus avoiding unwelcome surprises following the implementation.
Update any existing network diagrams you’ll be using. Be sure to label it so you know it is the original
(pre-IP telephony). Next, sketch your new network diagram with the gear included. Determine if there
is any overlap and if perhaps you don’t need as many switches as you thought. If you’re not working
with an integration partner, you may want to invest some money in having a technology expert take a
look at your new proposed network diagram. It’s better to make major changes in the planning stage as
opposed to after you’ve taken delivery of your IP telephony equipment. An expert can also make sure
you maximize your equipment purchase and may make modifications to your diagram that will save you
money in the long run.
After you’ve come up with your new network diagram, begin deploying the gear onto a test network.
This will not only help ensure the new system works optimally, it will help you get accustomed to the
new equipment so other deployments (to other locations, for instance) go smoothly. At the beginning,
the test network should not affect anybody’s workday. During the second phase, transition some non-
critical employees or departments to the test network. This will help you further test the system in a
real-world scenario and also gets users familiar with it.
Make sure that the team you’ve put together is available for the duration (right through user training), at
least on some level. If you’ve chosen a project leader, this is the person who will know all the details, even if
he or she is not working daily on all of them. Once you’ve made the switch, so to speak, sit back and start
enjoying the benefits of IP telephony.
between failure, mean time to repair, network reliability, and application reliability.
Ensuring Reliability
in IP Telephony 4
Chapter 4
Network Reliability
Application Reliability
ii
The most crucial characteristic of a business phone system is reliability. You must pick up the phone to
a dial tone, you must be able to successfully place outgoing calls, and calls must effectively reach your
organization. This chapter covers varying IP telephony solution architectures, mean time between
failure, mean time to repair, network reliability, and application reliability. It is meant to help you dig
deeper into the solutions you’ve narrowed down to your short list so that you can choose the one that
fits best into your organization and existing infrastructure and provide you with maximum uptime.
Availability, on the other hand, is predicted based on the probability of a hardware component failure. It
is predicted by taking into account the type and number of hardware components in a system and
calculating the mean time between failure (MTBF). So, if an IP switch has a predicted MTBF of ap-
proximately 135,600 hours, and each failure requires one (1) hour of mean time to repair (MTTR), we
would use this simple computation to estimate the availability:
This demonstrates that this particular unit will achieve “five-nines of availability.” Alternatively, this
switch is predicted to be unavailable for one hour every 10 years.
Let’s take a household example. Consider a toaster that works for a year (an average year is 365.2425
days = 8,765.82 hours or 8,766 hours), and then it breaks, so you have to replace it: MTBF = one year.
You take it to the store for a replacement the next day: MTTR = 24 (one day).
This indicates two-nines availability. However, if you keep an extra toaster on hand, MTTR could be as
little as fifteen minutes (.25 hours). While this increases the cost of equipment, it also increases the
availability fairly significantly.
Even repairing the problem in 4 hours doesn’t make it much easier to accomplish:
99.9990% = 400,000
400,000 + 4
You would still need 400,000 hours between failures. These examples are far beyond state-of-the-art. The
way to meet these demands is via redundancy. Read on for a section on redundancy and specifically n+1
redundancy.
A classic chassis includes a number of circuit boards, with most of them providing telephony interfaces and
one consisting of a specialized computer system, while some modular units contain a single board. The
classic chassis can be compared to a string of holiday lights: If one bulb fails, the entire segment fails. The
more lights on the string (number of circuit boards in the chassis), the more vulnerable it becomes to
failures.
A typical chassis model, because you have to take into consideration the reliability of their components,
typically has an MTBF in the 50,000 range, which is four (not five) nines availability. This can be raised to
five-nines by adding switches for redundancy (costly but effective). More on this will be discussed in the
n+1 redundancy section later.
Time
Be sure and ask vendors about their failure rates and how long a product lasts before end of life. If a
vendor does not give you a concrete number based on scientific calculations (not marketing hype), ask
more questions or talk to someone at the organization who can give you that information.
A 4-hour MTTR is industry standard, which creates a problem for IP vendors that want to maintain five-
nines of availability with a 4-hour MTTR. Redundant systems are usually added to ensure this availability
because a 4-hour MTTR requires a 400,000-hour MTBF to achieve 99.999% availability. (Availability =
MTBF/(MTBF+MTTR) = 400,000/(400,000+4) = 99.999%.) Modular, distributed systems tend to make
system repair easy, which results in a lower MTTR. These systems only require one power source and two
or three cable connections.
Redundancy also impacts the failure rate, ironically. While vendors often add redundant parts, such as disc
drives and power supplies, to their systems, the very fact that the number of parts are being doubled in
itself can increase the chance that the system will fail (increase the MTBF). When you are considering an
IP PBX system for your organization, be sure to look at how complex each system is. The more complex,
the longer it takes to repair because problem diagnosis, part replacement, and system restoration can be
difficult. Look for modular systems that are easy to manage and troubleshoot, with specific built-in tools to
ensure quick and easy diagnosis and repair.
In addition to a distributed architecture that provides n+1 redundancy, look for a solution that intercon-
nects each module using IP rather than cards in box slots. This design uses the Internet as a bus rather
than having a proprietary backplane, which allows you to use a wide variety of chips and software and also
reduces the costs and increases speed because of the use of IP and Ethernet. This design also allows you to
seamlessly scale your system to meet organizational growth demands, just as the Internet allows for growth.
Finally, look for a system that provides most of its feature upgrades via software so that there is minimal
time between the release and your organization’s use of these features.
Availability of N+1 system = 99.9 999 999 92%, that's "10-nines" or 4 million years
The goal of five-nines reliability is impossible for most systems because redundancy requirements can
be complex and expensive. Using n+1 redundancy is not only more cost-effective, but it is less complex,
which in turn reduces the chance of failure.
Network Reliability
The biggest hurdle when implementing an IP telephony solution is ensuring it works properly with the
existing underlying infrastructure. LANs and WANs have lower reliability than telecommunications
systems and are prone to quality-of-service (QoS) issues that make IP telephony solutions unreliable.
LANs have multiple serial components, which negatively affects the reliability (typical LANs achieve
three to four nines of availability), but it is possible to achieve five-nines availability on a network by
using a redundant aggregation switch with redundant paths. After all, four-nines reliability translates to
two hours of downtime per year. Can your organization afford that? Most 24/7 operations cannot. Focus
on solutions that allow these redundant paths to an aggregation switch.
A system with centralized call control relies heavily on its WAN connection because when it goes down,
remote sites have no call control, which means calls cannot be made unless a backup system is in place.
Look for a distributed solution that provides full and seamless call control functionality even during a WAN
failure.
Application Reliability
In addition to ensuring your system is reliable in terms of hardware, you must also ensure that IP telephony
system applications, including auto-attendants, voice mail, and desktop integration, work all the time for
your employees. Look at systems that offer one application server for a full range of applications. You can
use more than one server depending on your organization size, but make sure that it is not one feature per
server, like some solutions may force you to do. A truly reliable system, in terms of applications, uses a site
hierarchy, which means the first application in a user’s hierarchy is used, and each application server has
access to the configuration database in a central server. This design is highly reliable because each applica-
tion server caches the configuration database, making information and applications available even during
network downtime. For example, in the case of a network outage, remote users with their own server are
unaffected by a failure in another server so that individual sites can serve features like auto-attendant.
Handsets and
Interfaces 5
Chapter 5
The Need 1
Ergonomics 1
Sound 1
Screen Interface 2
User Considerations 2
Keypad Functionality 3
Soft Keys 3
Aesthetics 4
Phone Choices 4
Analog phones 4
IP phones 4
Soft phones 4
WiFi phones 5
Want a SIP? 5
American Disabilities Act (ADA) Compliance 5
ii
Think about your home telephone, your cell phone, or any one of the multiple electronic devices you
use everyday. You expect—and appreciate—a well-designed product. You shop for these items with
design and functionality in mind. With IP telephony, you can now bring the same high expectations into
the office and into your search for handsets. Mediocre office telephones are a thing of the past because
IP telephone handsets introduce so many more features and benefits.
The Need
Business workers rely on the telephone many hours out of the day, from collaborating with business
partners and co-workers to interacting with and helping customers and suppliers. Call center profes-
sionals literally spend the entire day on their telephones. It’s not enough to “make do” with a standard,
feature-lacking desktop handset. To make employees more productive—and happier—you need to
provide them with the tools they need to do their jobs optimally. You’ll only do this when you present
them with a handset that is ergonomically well-designed, has great sound quality, and features a multi-
tude of capabilities at the touch of a button.
Ergonomics
Ergonomics is the science of designing products, machines and systems to maximize the safety, comfort
and efficiency of the people who use them. Ergonomics takes into account psychology, physical mea-
surement, environment, and more to ensure that products are adapted to suit workers and their spe-
cific needs. Keep ergonomics in mind as you look at the handsets and graphical user interfaces (GUIs)
of each vendor’s solutions. If your organization is a machine shop, the most important feature for your
handsets may be a very loud ringer. If you have a call center staff, a bevy of features that help shorten
the call cycle will be most beneficial. A law firm may require a system that logs incoming and outgoing
calls and keeps this information on record for future reference. A recording studio may require ultra-
clear sound quality to ensure recorded voices are pitch-perfect. Look at your organizational needs in
terms of what you need a handset and GUI to do for your employees.
Sound
IP telephony, with its packet-based design, is able to deliver better than toll-quality sound with hi-
fidelity audio and innovative design. Better sound translates into productivity gains – shorter calls with
fewer errors, increased sales because of the clarity of conversation between a sales person and custom-
er, and increased caller satisfaction. Wideband audio is preferable over narrowband, because it has an
increased range on the low end (50-300 Hz) and makes conversations sound less tinny and reduces
error in translation. Look for a solution that supports both wideband and narrowband.
Figure 5.1 Wideband audio technology provides bandwidth from 50Hz to 7,000 Hz; narrowband provides 300 Hz to
3,400 Hz. Wideband delivers superior speech quality.
Speakerphone microphones are also an important part of sound quality consideration. Look for a solution
that supports hi-fidelity sound and has a full-duplex operation speakerphone so audio flows freely on both
ends (no delay if one speaker talks over another). Not all IP speakerphones are able to do this. In addition,
ensure you choose a handset that meets the Americans with Disabilities Act (ADA) regulations for the
hearing impaired, regardless of whether you have an immediate need or not. (More on ADA compliance will
be covered later in this chapter.)
Screen Interface
IP telephones act more like computers than telephones—they have a bigger screen and more functionality
attached to the screen. This screen also delivers more information about each call and prompts the user
through the call with various options appearing on the screen. The user simply presses a corresponding key
below the screen to accomplish any task while on the call (call forward, conference, etc.).
Make sure to consider carefully the size of the screen, with your users in mind. Is it big enough that after a
long day of work, it’s still pleasing to the eye? Is the display big and bright enough to see clearly after four
hours on the phone? In addition, work with the phone and test what features are available and how easy
those features are to access for a call center worker taking up to 50 calls an hour. Is there a message waiting
light to ensure no message is missed?
User Considerations
Another characteristic to consider is the feel of the phone, since that is another source of fatigue for users.
The phone should minimize shoulder and neck pain and fatigue, and it should essentially fit most users
comfortably. The handset should not be too light or too heavy—try and get a phone with a balanced weight
of about 170-190 grams. Also, consider a handset with a grip that is covered with a smooth rubber material,
as opposed to the slippery plastic kind that can become uncomfortable during long telephone calls.
• Directory: This key should be linked to a quick-dial program that allows a caller to dial by name
using the telephone keypad (7 for S, 2 for A, 6 for M, which would bring up names that match
beginning letters “SAM”).
• Redial: This function key should do more than simply dial the last number dialed—it should allow
you to press it and see an historical list of outbound, inbound and missed calls.
• Personal options: This feature key should allow for easy management of personal options, such as
ring tone and call handling preferences.
• Voice mail: This key should provide quick and easy access to voice mail messages.
Soft Keys
Soft keys are multi-function keys that use part of the telephone display to identify their function at any
moment. They are usually located directly underneath the display and their use changes depending on
where the user is in the call process. You can set some soft keys for use by all of your employees, and
you can choose to leave some to the discretion of each user. Make sure the setup of soft keys is straight-
forward before allowing users to set up their own. If the IP telephony system you’ve chosen does not
offer handset soft keys that are easy to set up or change, make sure the solution allows you to either set
the soft keys for each user (or block users from trying to set up their own) or the ability to choose not
to use the soft keys at all. This will minimize user confusion and frustration if the solution is difficult to
edit.
Easy to Manage
You want to make sure the phones you are getting with the IP telephony system you choose are plug-
and-play, particularly if you have a large organization with many locations, some of which have no
technical staff on-hand for installation support. Non-technical employees should be able to plug in their
phone and start working. When it’s plugged in, the phone should automatically get its IP address,
subnet mask, and gateway, as well as the accurate time from a time server. Handset updates should be
equally as hands-off for employees—updates should be automatic as they are released by the vendor.
Phone Choices
In an IP telephony solution, the IP-PBX manages telephones throughout the enterprise and acts as a
gateway to both voice and data networks. Any kind of telephone, whether it be analog, IP or a soft phone,
can connect to the IP-PBX via the network and calls are routed via the network instead of the public
switched telephone network.
Analog phones
A regular analog telephone, the same ones you’ve been using throughout your organization until now, can
be used in an IP telephony solution to input the caller’s voice into the system. Once in the system, a series
of analog-to-digital conversions and other processes change the voice signals into data, which is then
transmitted over the LAN, WAN, or Internet. The voice data is then converted back into sound by the
recipient’s phone. Most IP telephony systems will allow you to use your existing analog telephones with the
solution—forever or until you are able to afford and/or replace them with IP telephones. Be sure that your
vendor will allow you to phase out older analog phones with their IP phones over time so you can maximize
your existing equipment.
IP phones
IP telephones (or IP endpoints) actually perform the analog-to-digital and/or digital-to-analog conversions
and can plug directly into the LAN or WAN. VoIP system vendors usually offer a variety of IP telephones so
that you can choose different models based on various segments in your user population. Your legal depart-
ment may need multi-line handsets with easy conference call capabilities. A manufacturing floor needs a
phone with fewer bells and whistles but good, loud sound and a rugged exterior. Receptionists need hand-
sets with many more fixed feature buttons so that they can handle calls quickly and accurately.
Soft phones
A soft phone is essentially software that is used to make calls over an IP telephony system using a personal
desktop computer and either a headset connected to the computer’s sound card, or a telephone connected
to the computer using an adapter. It behaves like a traditional phone but usually offers much more informa-
tion to the user, depending on the vendor’s GUI. When a call comes into a station with a softphone, an icon
appears on the computer screen, which allows the user to either answer it by clicking on an icon, or ignore
the call by clicking on another icon, which in turn sends the caller to either voice mail or another employee.
Often, vendors offer an application that allows traveling employees to gain access to the robust feature set
of their desktop computer from wherever they are working—at home or on the road. A user simply logs into
the system from the local phone and has access to all of the same functions he or she would enjoy while in
the office.
One drawback to WiFi phones is the fact that some things can impede on the quality of the calls, such
as how many people are using the same hot spot, how close the WiFi phone user is to the access points,
WiFi card capabilities, and possible obstructions to the AP (such as a wall). Another drawback is that
WiFi technology does not offer the level of security offered with standard Internet access. More on
security will be covered in the following chapter.
Want a SIP?
Session Initiation Protocol (SIP), a signaling protocol, is used for establishing a session in an IP net-
work—from a simple two-way telephone call to a multi-media conference call session with many partici-
pants. The IP telephony industry has recently adopted SIP, an RFC standard (RFC 3261) from the
Internet Engineering Task Force (IETF), as the protocol of choice for signaling because of its ability to
facilitate Internet applications by working with other protocols. It is not the be-all and end-all of proto-
cols—it was designed to be a facilitation mechanism, not an all-inclusive solution. Its flexibility is what
makes it so powerful, and an all-inclusive approach does not offer this level of flexibility.
Essentially, SIP establishes, manipulates and tears down sessions, and its main purpose is to help
session originators deliver invitations to potential session participants wherever they may be. It uses
URLs to address participants and SDP to convey session information and it’s easy to combine SIP with
other applications, like Web browsers and messaging. The bottom line is that it’s a modular approach to
maximizing IP telephony protocols. SIP can find and invite call invitees wherever they are. It facilitates
multi-media calls with many participants who may join and leave at will.
• Volume Control: Telephones should have volume controls that provide a gain adjustable up to a
minimum of 20 dB. The telephones should provide at least one intermediate step of 12 dB for
incremental volume control.
• Automatic Volume Reset: The telephone should automatically reset the volume to the default
level after every use.
• Hearing Aid Compatibility: The telephone must have a means for effective magnetic wireless
coupling to hearing technologies.
• Minimized Interference: Interference to hearing technologies, including hearing aids, cochlear
implants, and assistive listening devices, shall be reduced to the lowest possible level that allows a
user of hearing technologies to use the telephone.
Security
6
Chapter 6
ii
Anybody who’s connected to the Internet or who owns a PDA/multi-function cell phone knows that
they’re at risk of getting viruses, worms, spam and other malicious threats. In addition to the potential
damage these threats introduce in terms of lost data or corrupted files, there are now regulatory issues
associated with ensuring protection. Healthcare has its own privacy regulations in the form of HIPAA
(Health Insurance Portability and Accountability Act of 1996), and infringements can result in signifi-
cant punishments and fines. The bottom line is that you have to protect your organization’s devices and
network. IP telephony is no different – the only difference is the form of the traffic: voice versus data.
All traffic crossing a network can be stolen, manipulated or blocked if proper network security precau-
tions are not put into place. This chapter will highlight the steps you should take to ensure your IP
telephony traffic is secure against outsiders and unauthorized individuals.
Phone Service Theft. A hacker could enter into an unprotected network and access the PBX to make
endless international calls. There have been major cases cited in the news where toll fraud has cost
companies millions of dollars. In many instances, the criminals have been caught and prosecuted, but
not without major costs to the companies defrauded; and keep in mind, there are always those crimes
that go undetected.
Eavesdropping. Without the proper security in place, a hacker could eavesdrop and possibly expose
confidential information. A private conversation about financials could be recorded and played for
anybody, which could lead to internal and external problems, including punishment from numerous
regulatory agencies. Or a personal call from an employee to a florist with a credit card number could
lead to credit card and even identity theft.
Power Failures. While outages affect data traffic, of course, there’s a difference when it comes to
telephony. People expect telephones to work even during an outage because homes often have a non-
electronic phone that simply plugs into the telephone outlet. This expectation is generally brought into
the workplace.
| Chapter 6 | Security
SPIT. Spam over Internet telephony is an alternative to telemarketing where one message can easily be
sent to thousands of recipients with the click of a mouse. In other words, your employees’ voice mail boxes
can become as overloaded with spam as their e-mail would be without appropriate spam filters.
Other Threats. There are new threats created and discovered daily. One such attack is the spoofing of a
phone number, which essentially allows a hacker to look like he or she is someone else, which is one of the
easiest ways for this person to steal an unsuspecting person’s identity. While individuals have learned not to
trust e-mail, it is still generally believed that telephone communications can be trusted.
• Physical security: Buildings, equipment rooms, data servers, and wiring closets should be off-limits to
anybody who is not authorized.
• Human security via security policies: Make sure your organization’s informational assets are protected
against inappropriate or unauthorized use by a renegade employee. Ensure hiring and system usage
policies are in place to govern appropriate use. Establish and strictly enforce policies having to do
with passwords and system usage.
• Network security: Again, create a multi-layered defense using firewalls, VPNs, and intrusion detection
or prevention (IDS/IPS). Make sure wireless access points use the highest level of access control and
encryption to prevent intruders from gaining access to your network and its resources.
• System security: Arm every desktop with anti-virus software to fight against spyware and other
malware. Utilize host intrusion prevention systems to protect servers against attacks.
Another force to consider is segregating traffic via virtual LANs (VLANs). It is a method of logically group-
ing devices or departments onto their own LANs. Isolating LANs from one another provides an additional
layer of security. It also reduces the impact of multicast or broadcast traffic since there are separate broad-
cast domains.
Finally, bandwidth management can be utilized to further guarantee bandwidth for business-critical,
latency-sensitive traffic like VoIP traffic. Bandwidth management methods include assigning a certain
priority to each type of traffic. VoIP packets should be assigned the highest priority to ensure voice traffic
gets through.
• Firewalls: Make sure the firewalls you’re using can handle the latency sensitive needs of IP telephony traffic.
• Switched environment: Use Ethernet switches (not hubs) to connect all your voice devices not only
for better performance but also to limit the possibility of a hacker getting onto a call because in a
Another architectural consideration to keep in mind is ensuring your system is distributed, which will
mean it has no single point of failure. A distributed system allows continued operation in the case of
worms, viruses, or DoS attacks. An attack will not disable the entire system if intelligence is distributed
amongst multiple devices.
Your chosen system should offer multiple levels for administrator permissions to limit control and
ensure unauthorized individuals do not gain access. Once you’ve deployed, reserve full access for just a
few key information technology employees. Ensure that a web-based management solution supports
secure management using Secure Sockets Layer (SSL), which secures communications from the
interface to the server.
According to the SANS (SysAdmin, Audit, Network, Security) Institute, a cooperative research and
education organization, VoIP servers and phones are at significant security risk. The organization’s 2006
annual update, SANS Top-20 Internet Security Attack Targets, indicates that there’s been an increase in
security scrutiny of IP telephony, especially on typical components such as the call proxy and media
servers, as well as the phones themselves. Some products have been found to contain vulnerabilities
that can either lead to a crash or a complete control over the server or device. “By gaining a control
over the VoIP server and phones, an attacker could carry out phishing scams, eavesdropping, toll fraud
or denial-of-service attacks.”
| Chapter 6 | Security
How to Mitigate IP Telephony Vulnerabilities
SANS has determined and published a list of things enterprises must do to mitigate the IP
telephony vulnerabilities mentioned in this chapter.
• Apply the vendor supplied patches for VoIP servers and phone software/firmware.
• Ensure that the operating system running the VoIP server is patched with the latest OS
patch supplied by either the OS vendor or the VoIP product vendor.
• Scan VoIP servers and phones to detect open ports. Firewall all ports from the Internet that
are not required for keeping up the VoIP infrastructure.
• Use a VoIP protocol aware firewall or Intrusion Prevention product to ensure that all UDP
ports on VoIP phones are not open to the Internet for RTP/RTCP communications.
• Disable all the unnecessary services on phones and servers (telnet, HTTP etc.).
• Use VoIP “protocol fuzzing tools” such as OULU SIP PROTOS Suite against the VoIP
components to ensure the VoIP protocol stack integrity.
• Additional caution should be taken at the product selection phase to ensure the VoIP
product vendor supports OS patches as they are released. Many VoIP vendors will void
support for unapproved patches and may take considerable time before approving them.
• Apply separate VLANs to your voice and data network as much as your converged network
will allow. Ensure that VoIP DHCP and TFTP servers are separate from your data network.
• Change the default passwords on phones’ and proxies’ administrative login functions.
Source: SANS Top-20 Internet Security Attack Targets, 2006 Annual Update
Going Mobile 1
Wireless Next 2
Sufficient Coverage 2
Scalability 3
Quality of Service 3
Seamless Roaming 3
Solid Security 3
Selecting Handsets 3
ii
In addition to cost savings, productivity improvements, and customer service enhancements, another
driving force behind IP telephony is mobility. Workers are increasingly mobile—from traveling sales
people to call center staffers who work from remote sites and even home offices around the globe to
serve customers 24/7. Mobility is an absolute necessity, as is the requirement for customers to reach
anyone at anytime, anywhere. IP telephony is the ideal way to meet this need. With it, organizations can
use distributed hunt groups to ring employees around the globe with the right skill set to ensure a
question is answered or an issue is solved immediately. As long as an agent with the skill set is logged
in, even if on another continent, the issue will be resolved just as if he or she were at their own desk.
With IP telephony, calls are intelligently routed based on calendars, so agents logged as out of the office
are reached via cell phone, etc. At the same time, the agent’s cell phone acts as an extension of their
desk phone with all of the integrated features, such as dial-by-name, transfer, conference call capabili-
ties, etc. This wireless integration is crucial, especially since you don’t necessarily need to purchase
specific wireless handsets or specialized handsets for traveling employees.
This mobility is not even noticeable to the customer base. There are IP telephony vendors that allow
employees to choose their device—for instance, a cell phone or home phone—and that device assumes
the identity and capabilities of his or her regular office extension. For example, the caller-ID informa-
tion provided when the employee makes a call can reflect their office number instead of the mobile or
home-office phone actually being used. In other words, caller-ID will indicate that the call is coming
from headquarters of their company. This is important to protect the employee’s privacy and strengthen
the corporate brand.
Going Mobile
With IP telephony, users are highly mobile, logging in from anywhere and gaining access to all the same
capabilities as if they were working at headquarters, at their desks, or within a call center building. With
IP telephony, to the outside world, it can seem as though your organization has call center locations
scattered around the globe, making help available 24/7. In reality, you are simply utilizing IP telephony
features such as time-of-day routing and call forwarding to make sure calls are answered quickly by a
live human being. Your employees can be working out of branch offices, at remote locations, or even at
home. Your workers are mobile and happy; your customers are being catered to and satisfied quickly.
You are also able to manage peak calling times by having the ability to add other employees, regardless
of their location, to the call center to help meet the overflow demand.
With IP telephony, users can also easily re-route their calls so that they are reached wherever they will
be working—they can make these changes themselves, without asking for IT assistance. This “find me”
feature also enhances customer service, as well as productivity, by ensuring every call reaches the right
person, regardless of where he or she might be working. An employee can even program his or her
extension to ring based on status—ring through when he or she is in the office, forward to a cell phone
when there is no answer, or forward to a colleague when the line is busy.
With wireless IP telephony, employees are not tied to their desks and delays are further reduced. Consider,
for instance, the case of the sales representative meeting with the CEO. While in a meeting, urgent calls can
follow him or her to a wireless handset. Take this example into a hospital, and it can mean the matter of life
and death if a nurse is visiting a patient whose health suddenly degrades. The nurse need not waste time
running to the nursing station to call the doctor or paging for help but rather, he or she can call the doctor
directly from within the patient’s room from a wireless IP handset, provide information and take steps the
doctor is advising all in real time as a result of the phone consult.
On top of the savings offered by IP telephony, going wireless can also save your organization additional
money. For example, when an employee is working in another location other than his or her office, calls can
still find that person if they are free to talk, thereby eliminating any toll charges that would have been
associated with returning a missed call, had a caller gone to voice mail.
Sufficient coverage
You don’t want your users hitting dead zones while they’re in the middle of a conversation. It’s poor customer
service and costly to your business. Assess how many users you have in each location of your organization,
and consider the bandwidth requirements of the applications they are each running to ensure enough band-
width for voice traffic over the WLAN. You will need to maximize performance by adding a sufficient number
of wireless access points (APs) to each location where many users work. Keep in mind that since a WLAN is a
radio frequency (RF) network, the physical environment will affect the coverage capabilities of each AP. Walls,
glass partitions, and cubicle separators can affect the coverage area because these materials absorb signals.
Take into account the physical characteristics of your organization and buildings and design your WLAN plan
to meet these challenges. A physical survey before deployment will help you determine how many APs and
switches you’ll need to meet coverage requirements. Keep in mind, however, that the more APs you add to a
particular area will affect performance in terms of possible interference.
Quality of Service
Delays for voice should not exceed 150 ms, and given that Wi-Fi is a contention protocol, when an
access point is overloaded, voice quality will suffer. QoS is required for voice traffic whether it’s travers-
ing a wired or a wireless network. In other words, you want QoS for your voice traffic over the air or
over land so look for gear that offers over-the-air quality of service. Guaranteeing voice over other
applications minimizes packet loss, delay and jitter that results in poor voice quality. The IEEE is
working on a standard to address QoS for wireless networks, but in the meantime, the Wi-Fi Alliance
has released Wireless MultiMedia (WMM) as a subset of these capabilities. Vendors are currently
bringing WMM implementations to market now. WMM defines four priority levels to support varying
kinds of traffic, including voice, video, best effort for data, and background traffic, in that order.
Seamless roaming
As a user walks from one office or location to another, he or she counts on roaming capabilities of the
WLAN to keep the call connected. The underlying wireless infrastructure must seamlessly hand off the
user to the next location and perform the necessary re-association and re-authentication with APs,
while keeping calls free of interruption (this will allow a call to continue seamlessly across zones
without being mistakenly dropped between zones). A security standard is under way to allow users to
be pre-authenticated to neighboring APs before roaming, which will reduce the time it takes for a user’s
call to move between APs, and in the meantime, some wireless equipment vendors are introducing their
own versions of fast-roaming capabilities.
Solid security
IEEE 802.1X authentication should be used to verify a user’s identity onto the network, which will
ensure unauthorized guests are not allowed entrance to use the network or gain access to confidential
corporate information. Laptops and handhelds can support 802.1X authentication, and you need to
make sure your wireless IP phones, which have less computational capacity, are using less processor-
intensive authentication methods like MAC address or username and password.
Selecting Handsets
As discussed in Chapter 5, Session Initiation Protocol (SIP), a signaling protocol, is used for establish-
ing a session in an IP network—from a simple two-way telephone call to a multi-media conference call
session with many participants. The VoIP industry has recently adopted SIP, a RFC standard (RFC
3261) from the Internet Engineering Task Force (IETF), as the protocol of choice for signaling because
of its ability to facilitate Internet applications by working with other protocols. Essentially, SIP estab-
lishes, manipulates and tears down sessions, and its main purpose is to help session originators deliver
It is possible to use traditional cell phones and they can become an extension of your IP telephony solu-
tion—this requires no new wireless network or SIP handsets. There are also many wireless IP telephony
handset vendors out there, but they don’t all offer the same features. Start your search by looking at your
needs first. Are you looking at wireless handsets for a manufacturing facility and therefore need a rugged
handset with dust covers so they don’t get dirt inside the keys? Are you a healthcare facility and need to
meet safety requirements so the handsets don’t interfere with hospital equipment? After you determine
your general needs, next move on to what you would like to see the handsets offer. Would you like the
handsets to be able to transfer calls? Would you like your employees to be able to conduct conference calls
from the wireless handsets? What’s your wish list on top of your needs list? These two things will bring you
to a number of vendors’ solutions, and then the final question you need to ask is, will it work with your IP
PBX vendor’s solution? Your choice will be very easy at this point—you’ll likely either have just one or two
vendors left from your list.
Quality of Service
8
Chapter 8
Quality of Service 1
Leased lines 2
Frame Relay 2
MPLS 2
Internet VPN 3
Identification Methods 5
DiffServ or ToS 5
802.1p 5
VLANs 6
Prioritization Methods 6
Priority Queuing 6
Custom Queuing 6
ii
Quality of Service
As we drill down deeper into the details about your converged network, going over topics like handsets,
security, and mobility software, you are most likely growing more comfortable with IP telephony.
However, you probably still have some concerns, most notably, “How do I know my boss isn’t going to
experience poor audio quality during calls?” or “How can I be certain that all of our average 200 calls
are always going to get through, even on our busiest day, and when accounting is doing its weekly
check run?”
The answer is Quality of Service, or QoS. This chapter will cover QoS in detail, as well as your options in
terms of circuit transports, and then delve into the internal infrastructure and the entire process of
applying QoS. QoS can be boiled down to three major steps: Identify. Classify. Prioritize.
You need to consider four things that can affect voice traffic:
1. Latency (or packet delivery delay )
2. Jitter (or the variation in time between packets)
3. Packet loss (which can occur when too much traffic overflows buffers within the network causing
packets to be dropped), and
4. Burstiness (when your network undergoes bursts of packet drops due to jitter)
It’s bad business to have your voice traffic burdened by any of these effects. Distance alone on the WAN
circuit can cause delay, as can lower-speed WAN circuits. Delays cause call participants to start inter-
rupting each other because they believe the other person is finished speaking. Latency should not
exceed 100 milliseconds (ms) one way for toll-quality voice and must not exceed 150 ms one way for
acceptable quality voice. At 150 ms, delays are noticeable by the human ear, but callers can still carry
on a normal, comfortable conversation.
Jitter can cause strange sound artifacts to contaminate the voice and users will complain of degraded
voice quality. Jitter has many sources: network congestion, queuing methods used in routers and
switches, or network routing policies such as traffic engineering or MPLS paths used by carriers.
If your phone conversations do not sound right and callers have to keep repeating themselves or have a
less than satisfactory experience when they call, they’ll start looking for other ways to communicate
with your employees, or worse, they’ll start looking to another company to serve their needs—one with
which they can communicate more clearly. Your IP telephony system should sound better than your
previous phone system—after all, that’s why you made the switch. It’s the only way you’ll ensure that
you don’t lose business because of your technology change. IP telephony should increase—not de-
crease—your business and your bottom line.
Leased lines
Leased lines are the most private way to go. They are also the easiest type of WAN circuits to configure
guaranteed QoS. These circuits are direct point-to-point lines connecting your locations together. They can
be used for data, including packetized VoIP, or Internet services.
Frame Relay
Frame Relay circuits are more economical than private leased lines because the Telco providing the Frame
service shares bandwidth among many subscribers. This can reduce your costs, especially for long distance
lines, but commonly reduces your guaranteed bandwidth to less than your full circuit speed. Frame Relay
can guarantee bandwidth and packet delivery only if you shape your outgoing traffic to match your com-
mitted information rate (CIR). Properly engineered, Frame Relay can provide a cost-effective means of
transmitting IP telephony traffic and still guaranteeing QoS.
MPLS
MPLS, like Frame Relay, is a label-switched system that can carry multiple network layer protocols. Similar
to Frame Relay, MPLS sends information over a wide area network (WAN) in frames or packets. Each
frame/packet is labeled and the network uses the label to decide the destination of the frame.
MPLS header
Label Exp S=0 TTL Label Exp S=0 TTL Label Exp S=1 TTL IP header TCP header Payload
Figure 8.1 | MPLS works by pre-pending packets with an MPLS header, containing one or
more ‘labels’. This is called a label stack. Source: Wikimedia Commons.
MPLS networks can use Frame Relay, ATM or leased lines for the link layer.
Internet VPN
A virtual private network (VPN) is a private network used by an organization or in many cases by a
company and its partners or associates, to communicate or coordinate confidentially over a non-private
network. VPN traffic can be carried over a public networking infrastructure such as the Internet.
Internet-based VPNs offer the least amount of administrative control to regulate and guarantee QoS.
Your spectrum of options runs from leased lines combined with feature-rich switches and routers at one
end of the spectrum, to Internet-based VPNs using consumer-grade WAN circuits (DSL, Cable-mo-
dems) from separate providers with no SLA.
With the first option, you have complete administrative control over all points of congestion and have
the configuration tools and features to easily identify, classify and prioritize your VoIP traffic. This is the
optimal choice if you have the budget for it. Managed routers with features that you cannot control and
circuits that you do not have administration control over can be less effective for your network and
often require more labor to ensure configurations are correct and guarantees are being followed by the
managed service provider. You have less power in terms of making forwarding decisions and changes on
the fly. At the other end of the spectrum, you have cast off all control over every components, circuit
and congestion point and have thrown your VoIP packets into the Internet with simply the hope that
they get there, but effectively powerless to help them arrive safely and on time.
Minimizing Latency
Latency (also known as delay) is the time that it takes a packet to make its way through the network to its
destination (or the time it takes the speaker’s voice to reach the listener’s ear). Actually, some latency is
inherent and constant due to distance and the number of devices in the path. As mentioned, large
latency values can cause hesitations and, therefore, call participants interrupting one another. There can
be a number of factors that contribute to latency, such as propagation delays (the time it takes an
• The faster the media, the less time it takes to serialize the digital data onto the physical links, and
the lower the overall latency. The impact on latency depends somewhat on the link technology used
and its access method. For example, it takes 125 microseconds to place one byte on a 64Kb circuit.
Placing the same byte on an OC-3/STM-1 circuit takes 0.05 microseconds.
• Although some delay is unavoidable regardless of the bandwidth used, keeping the number of
intervening links small and using high bandwidth interfaces reduces the overall latency.
• The packet forwarding delay is determined by the time it takes a router, switch, firewall or other
network device to buffer a packet and make the forwarding decision. Among the forwarding consider-
ations are which interface to forward the packet to and whether to drop or forward the packet against
an Access Control List (ACL) or security policy. Packet forwarding delay varies depending on the
function and architecture of the networking device. If a packet must be further buffered as a part of its
processing, greater latency is incurred. (Source: VoIP 101, Juniper Networks.)
Time
20 ms 20 ms 20 ms
32 ms 8 ms 25 ms
Jitter is caused by congestion or other factors. Most media gateways have play-out buffers that buffer a
packet stream, so that the reconstructed voice stream is not affected. Play-out buffers can minimize the
effects of jitter, but cannot eliminate severe jitter. Although some amount of jitter is to be expected,
severe jitter can cause voice quality issues because the media gateway might discard packets arriving out
of order. In this condition, the media gateway could starve its play-out buffer and cause gaps in the
reconstructed waveform. (Source: VoIP 101, Juniper Networks.)
Once each high-priority packet has been marked with your corporate standard (for instance, DiffServ),
then at egress (as the packet leaves a piece of networking gear such as an Ethernet switch or router), it
needs to be prioritized above other packets. Keep in mind there are different levels of priority as well as
different queuing methods, so if your organization is a hospital, you will likely have healthcare applica-
tions, such as patient records and networked images transfer applications, assigned a higher priority
along with voice traffic.
Identification Methods
The following is a list, although not exhaustive, of identification methods you may choose to utilize for
prioritizing voice traffic. Again, your IP telephony vendor may choose one method and you may choose
another for your corporate standard, in which case you will be re-marking each packet with your
method choice.
DiffServ or ToS
Layer 3 QoS using DiffServ or Type of Service (ToS) bits is a system of identifying IP packets by assign-
ing values within the layer 3 IP header. Once identified, traffic can be classified into groups so that QoS
policies can be applied. For example, maybe Web access needs to be reasonably responsive but accept-
able e-mail response time can range from seconds to minutes. On the other hand, voice traffic (IP
telephony) and IP videoconferencing require a much higher level of priority. The type of end-to-end
QoS you choose to implement will depend on what type your routers and IP telephony solution support.
DiffServ and ToS add state information to each packet—allowing the network equipment to identify
different service flows and direct queuing and forwarding treatment appropriate to the service require-
ments. This enables routers to identify voice packets and mark them for higher priority treatment over
less sensitive packets. With DiffServ or ToS, each router on the network is configured to differentiate
traffic based on its class and each traffic class can be managed differently, insuring preferential treat-
ment for higher-priority traffic on the network.
802.1p
802.1p is a specification that gives Layer 2 switches the ability to identify and prioritize traffic. It works
at the media access control (MAC) framing layer (Layer 2) of the OSI model. Eight classes are defined
by 802.1p, which uses the priority fields within the packet’s VLAN header to signal the switch of the
priority-handling requirements.
Prioritization Methods
Once you’ve decided how you’re going to tag your high-priority packets, next you have to determine your
prioritization method. Here are just a few.
Priority Queuing
Priority Queuing supports multiple fixed-length queues from high to low, servicing the highest queue first,
then the next-lowest priority and so on. If a lower-priority queue is being serviced and a packet enters a
higher queue, that queue is serviced immediately. While good for important traffic, it can lead to queue
starvation.
Custom Queuing
Custom Queuing is designed for environments that need to guarantee a minimal level of service to all
protocols. It allows a customer to reserve a percentage of bandwidth for specified protocols. Customers can
define multiple output queues for normal data and additional queues for system messages such as LAN
keepalive messages. Custom Queuing can guarantee that mission-critical data is always assigned a certain
percentage of the bandwidth, but also assures predictable throughput for other traffic. (Source: Custom
Queuing and Priority Output Queuing, Cisco)
Resources:
http://www.wikipedia.com
http://www.networkworld.com/links/Encyclopedia/index.html
http://www.juniper.net/solutions/literature/white_papers/200126.pdf
960 Stewart Drive Sunnyvale, CA 94085 USA Phone +1.408.331.3300 +1.877.80SHORE Fax +1.408.331.3333 www.shoretel.com
Copyright © 2007 ShoreTel. All rights reserved. ShoreTel, the ShoreTel Logo, ShoreCare, ShoreGear, ShoreWare, ShorePhone, ControlPoint and Office Anywhere are trademarks or registered trademarks of ShoreTel, Inc.
All other marks are the property of their respective owners. Specifications are subject to change without notice. Part #EB-01/4.07